Measurement and modelling of combustion in a spark ignition engine

Abstract

A study has been conducted into the causes of cycle by cycle variations in combustion within a spark ignition engine, the best measured engine parameter to use for its characterization, and the effects that: ignition timing, equivalence ratio, fuel type, throttle position and knock, have upon it.

A Ricardo E6 single cylinder variable compression ratio research engine was instrumented to allow measurement of: cylinder pressure, temperatures, speed, load, fuel flow and air flow. The engine was also fitted with an optical slice that allowed optical access to the combustion chamber and enabled measurement of the early flame speed (up to 10 mm from spark plug gap) using a laser schileren system. Cylinder pressure data were collected on a dedicated HP1000 computer for every degree of crank angle rotation for up to 300 successive cycles.

A phenomenological model was developed for turbulent combustion that split the combustion process into three phases: early laminar burn, turbulent combustion, and final burn. The model allowed the study of the physical phenomena occurring within the combustion chamber and enabled insights to be gained into their effects on combustion and cyclic variations.

The study showed:

The variation in mixture strength has a far greater effect on the average and Coefficient of Variation (COV) values of all the combustion performance parameters, than does changing the fuel type. Cycle by cycle variations in combustion are best characterized by COV of imep.

The onset of knock has no discernible effect on the COVs of the measured parameters. The part throttle results show higher COVs than at Wide Open Throttle (WOT) due to slower burn, supporting the theory that faster initial flame speeds reduce cyclic variations.

The combustion model was used to support the hypothesis that cycle by cycle variations are caused by movement of the flame kernel by turbulence within the combustion chamber.